MalwareClassifier

Requirements

this project uses jupyter notebooks, for tutorial on usage see jupyter-notebook
requirements.txt
- contains modules needed to execute the code which are not included in the standard python distribution (apart from the jupyter notebook)
- to install run: pip install --requirement requirements.txt

folder Scripts contains:
- TestDataExtraction and TrainDataExtraction subfolders.
  - Scripts within these folders are responsible for extraction of training and test data
- TrainFeaturesScaling
  - Scales all training samples into the range [0, 255]
- SingleFeatureProcessing
  - Contains overview of each feature together with a classifier based on each feature
- WholeFEatureSetClassification
  - Contains LeNet-5 and random forest classifier for the whole feature set
- utilities
  - contains helper classes and functions
folder Data.zip:
- contains only the final version of training and test features in feature_merged_scaled and Data/test_data/test_features_merged_scaled
- the whole dataset is not provided as it is several GB in size, but it can be found in the Ember dataset
- upon downloading the following version: ember_dataset_2018_2.tar.bz2 both training and test data can be extracted using corresponding extract_to_csv.ipynb jupyter notebook